Data can be the next oil for Africa, but you still need a place to store it
A World Economic Forum panel discussion broadcast from Cape Town in September 2019 talked at length about the enormous possibilities data has for African economies from a growth and development perspective.
Among various figures and projections offered by the four experts on the panel, the one that caught my attention was the mention that data’s contribution to the African economy as a whole, will probably be measured in trillions of dollars by about 2035. This sounded a bit over the top but since the time of the discussion, several things have happened so that I am now a believer when it comes to the potential data has for Africa.
At that time (2019), the world was in pre-lockdown mode. On the surface everything appeared normal and the 2008/09 great world recession was just about a decade ago, so no longer featuring on anybody’s radar too prominently.
Also noteworthy, all the items I have tracked that discuss the role of data in Africa’s economic development, focus on the data itself, in other words, data as a commodity. And I must state that the dreams and ideals for all the data that we are going to collect, were indeed lofty, but that was before Covid. I doubt the focus on the utility of data has managed to come through Covid unscathed.
And then this week, I received an article for publication that also punted very large figures for data, but this time the focus was not on the data itself, but on the infrastructure required to handle, store and process all this data. This opened entirely new vistas in my mind especially when I started reading a little about existing and proposed data centres.
Data, like any commodity, has to have a primary producer, let’s say the miner. Then it has to have value chains to reach the market, in today’s world that is the internet, and then there has to be a place where it can be stored so that all the clients who need it, can have access to it, and that is where a data centre comes in.
The proliferation in data over the past ten years is so astronomical, ordinary people battle to follow the terminology geeks use to describe the speed at which data travels, and the size of data storage facilities. So, by definition, data centres are huge, or more accurately, massive but not in footprint, only in capability and capacity. They are massive in terms of what one can store in them compared to the usual computing power that either sits on your desk in the form of a PC, or sits in a company’s server room where one very powerful, or several large computers, serve as the company’s data server.
Enter the concept of the Cloud. If there are such massive amounts of data looking for an abode, and if there are certain, tangible risks for any company to store its own data on-site so to speak, were it not better and safer to store the data in a dedicated facility cared for by teams of technicians. And given today’s internet speeds, does it not make more sense to send all a company’s data to such a safehouse, and then only use it when needed?
In a relatively short time, the Cloud has gained traction, growing rapidly in utility and popularity, to the extent that a very large proportion of all data business now happens in the Cloud.
But the Cloud also needs a place where it can hover and this is where Data Centres come in. For this magical, nefarious concept of the Cloud to be practical and to actually work, it needs storage capacity and internet connectivity at a level that truly boggles the mind. And the more data centres there are, the bigger, more stable and more secure the Cloud becomes.
So when I started reading the article about a company that already operates so-called internet exchanges and that now wants to build ten very large data centres across the continent, I figured, Why not in Namibia also?
Data centres are not very big physically. They can be but they do not need to be. Data centres are a relatively new development in IT so the jury is still out as to what is the optimal design. But I am fairly certain that within a few years, data centre design will have reached a certain standardised format.
Data centres are notoriously not-green. They consume huge amounts of electricity, but here is the catch, mostly to keep them cool. The amount of energy they need for computing is actually far less than what they need not to overheat.
Furthermore, they like stable environments with the absence of moisture and without any extremes in temperature. Where this is not possible, they require even more electricity to modify the air and to keep temperatures stable.
So why can we not erect data centres in the desert, and I mean dozens of them?
A Tier-IV megascale data centre costs around US$50 million, or about 750 million of our schmekels. That is not a huge amount of money so investment and financing are not insurmountable obstacles like with so many other technologies.
They can be built in the desert, close to where there is existing infrastructure, like Arandis, and they will obviously have to be close to the line that connects to the WACS undersea data cable. They can be built semi-undergound where temperatures are relatively stable, especially in the desert where nights can be cold.
For a roof, doubling as a sun-shade, the whole data centre can be covered by photovoltaic solar panels, generating the electricity they need by day, and charging the batteries on which they run by night.
And if the developer is really ambitious, the whole installation and its solar panels, can be turned into a smallish solar chimney like the one at Mantanarres in Spain, and continue generating its own electricity during the night.
There are many options and possibilities. The only thing we lack is a visionary Namibian who can tie all the ends together, get the investors, the contractors and the future clients together, and start building a battery of data centres.
What’s more, none of those clients need to be in Namibia so the small local market is not a deterrent. They can be anywhere on the continent, if the focus is Africa only. And perhaps the biggest advantage we must consider is the fact that data centres need armies of technicians to make sure they work properly and continue to work properly. This is the reason why I suggested a location near Arandis as a starting point, but they can be near any town that can be connected by fiber to the WACS runner.
We do not need to focus on data as a commodity to make it work for us, we can shift our gaze and focus on what makes it work for other people i.e. clients, and this way create enormous opportunities for ourselves.